Damage Identification of Thermal Protection Tiles Based on an Optimizable SVM Algorithm and Temperature Monitoring

During the service of reusable launch vehicles, thermal protection tiles are easily damaged, which in turn threatens flight safety. However, monitoring the back wall temperature of thermal protection tiles can reflect their damage status. Based on this principle, combined with machine learning methods, we proposed a thermal protection tile damage identification method based on an optimizable SVM algorithm and temperature monitoring. The method is verified by the bottom temperature data of the thermal protection tile collected from ground simulation tests. Results show that the optimizable SVM algorithm can quickly identify various types of thermal protection tile damage, with an identification accuracy of up to 98.44%. This work lays a foundation for further development of thermal protection structure health monitoring and management system.